基于纹理特征的恶意代码检测方法测试

Malware Detection Method Based on Texture Feature

为了有效避免恶意代码给网络信息安全带来的破坏,对恶意代码检测方法进行研究。在恶意代码可视化的理论基础上提出一种基于纹理特征的恶意代码检测方法,通过可视化算法实现对恶意代码二进制程序的可视化显示,即映射为灰度图像,再提取图像的纹理特征,并在恶意代码语料库中进行特征匹配,输出检测结果。最后利用在某公司安全部门捕获到的5种出现次数最多的恶意代码种类进行实际检测实验,实验结果表明,基于纹理特征的恶意代码检测方法实用性强,对恶意代码的分类较精确。

In order to effectively avoid the damage of malware to the network and information security, a malware detection method was investigated in this paper. In the light of the theory of malicious code visualization, a novel malware detection method based on the texture feature was proposed. The visualization algorithm is used to visually display the binary program of the malicious code, i.e. the gray-scale image. Then, the texture feature of the gray-scale image is extracted to matches the texture-feature in the library of malicious code and output the detected results. At last, 5 types of malicious code captured by the security department of a certain company, which appear most frequently, are practically tested. Experiment results show that, the proposed method has the strong practicability because it can precisely classify the malicious code.